Author's personal copy - Romana Cerc Korošec nickel oxide....pdf · Author's personal copy glass (square resistivity 13 /ð) as the working electrode. For background measurements,

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Electrochromic nickel oxide/hydroxide thin films prepared by alternatelydipping deposition

R. Cerc Korošec a,⁎, J. Šauta Ogorevc a, P. Draškovič a, G. Dražić b, P. Bukovec a

a Faculty of Chemistry and Chemical Technology, University of Ljubljana, Aškerčeva 5, SI-1000 Ljubljana, Sloveniab Jožef Stefan Institute, Jamova 39, SI-1001 Ljubljana, Slovenia

Received 1 April 2007; received in revised form 3 March 2008; accepted 11 March 2008Available online 19 March 2008

Abstract

Nickel oxide/hydroxide thin films were prepared by alternately dipping deposition from NiSO4 and LiOH solutions followed by heat-treatment.The onset temperature for nickel hydroxide decomposition was 220 °C for a thin film sample and 325 °C for the corresponding xerogel. During thedecomposition nickel oxide was formed. The evolution of the infrared spectra during thermal decomposition is explained in detail. The thermallyuntreated xerogel contained sulfate ions, whereas in the thin sample they were replaced by carbonate ions originating from atmospheric CO2.Optimization of the electrochromic response of thin films was performed using in situ spectroelectrochemical measurements. In the optimized filmthe degree of thermal decomposition of nickel hydroxide was 65% (120 min at 235 °C). The change in transmittance between the bleached andcoloured state was low at the beginning of cycling (13% at λ=480 nm), but increased during cycling and reached 45% in the 101st cycle (colourationefficiency −37 cm2 C−1). Transmission electron microscopy was used to examine the structure of the optimized film and of two more intensivelythermally treated films. One of them (exposed to 300 °C for 24 h) became electrochemically active during cycling while the other (24 h at 400 °C)remained inert.© 2008 Elsevier B.V. All rights reserved.

Keywords: Nickel oxide; Nickel hydroxide; Thin films; Electrochromism; Thermogravimetry; Alternately dipping deposition; Transmission electron microscopy;Optimisation

1. Introduction

Electrochromic nickel oxide thin films are good candidatesfor ion-storage materials in combination with tungsten oxide incomplementary electrochromic devices [1]. Under anodic po-tentials NiO films change colour from transparent to deep brown.For chemically prepared NiO thin films the magnitude of theoptical modulation during potential switching is crucially in-fluenced by their heat-treatment, which is always performed afterthe deposition process [2–5]. For sol–gel prepared NiO films ithas been shown that optimisation of thermal treatment on thebasis of thermogravimetric (TG) analysis of the films assureshigher coloration efficiency values, as well the stability of the filmstructure after prolonged cycling in an alkaline electrolyte [2,5,6].

Temperatures obtained for the corresponding xerogels are higherthan for thin film samples by approximately 30 °C due to thedifference in sample size. Therefore, optimisation and also expla-nation of the structural and morphological changes that occurduring thermal treatment should be made on the basis of dynamicand isothermal TG measurements of thin films deposited on asubstrate. The results show that each chemical system should beoptimised separately regarding the preparation route [6].

The aim of our work was to investigate the temperature de-pendent electrochromic response of Ni(OH)x thin films, preparedby the alternately dipping deposition (ADD) technique. Thermaldecomposition of the film deposited on a substrate was comparedto that of the xerogel. The difference ofmore than 100 °C betweenthese two samples in the temperature at which NiO begins to formwas explained with the help of Fourier transform infrared spec-troscopy (FT-IR) of thermally untreated samples and intermedi-ates formed during thermal decomposition. This study therefore

Available online at www.sciencedirect.com

Thin Solid Films 516 (2008) 8264–8271

⁎ Corresponding author.E-mail address: [email protected] (R. Cerc Korošec).

www.elsevier.com/locate/tsf

0040-6090/$ - see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.tsf.2008.03.014


upgrades and completes work where thin film properties wereexplained by TG and X-ray diffraction (XRD) measurements ofthe corresponding xerogel [7].

2. Experimental details

2.1. Preparation of thin films and xerogels

Thin films were prepared on different substrates by the ADDtechnique. First, a wetting agent was dispersed on the substratewith a pulling velocity of 5 cm min− 1. For platinum foil andSnO2/F glass, 1 wt.% of Teloxide (TEOL Factory, Ljubljana,Slovenia) in ethanol was used as the wetting agent, while forsilicon resins 1 wt.% of Etolat (TEOL) in distilled water was used.After the wetting solution had dried, thin films were prepared byalternately immersing the substrates in two solutions, the first of0.05 mol L− 1 NiSO4 (Kemika, Zagreb, Croatia) and the other of0.05 mol L− 1 LiOH (Kemika). Consequently a thin film of lightgreen amorphous nickel hydroxide was formed on the substrate.The thickness of the thin film increased with the number ofimmersions, as noticed by observation. For TG measurements,thin films were deposited on platinum foil. For FT-IR analysissilicon resins polished on both sides were used as substrates. Thenumber of immersions was 7 for each solution. For in situ spec-troelectrochemical measurements, SnO2/F glass was used as asubstrate. The number of immersions was again 7. The xerogelused in TG and XRD measurements was obtained by mixingequal volumes of 0.1 M NiSO4 and 0.1 M LiOH, centrifuged,cast in a Petri dish and dried in air at ambient temperature.

2.2. Instrumental

2.2.1. TG measurementsThermoanalytical measurements were performed on a Mettler

Toledo TG/SDTA 851e instrument. In dynamic measurements,the temperature range from room temperature up to 900 °C wasapplied for the xerogel and thin films deposited on platinum foil. Inisothermal measurements, the furnace was heated at 5° min− 1 tothe chosen temperature, left at that temperature for 180 min andthen heated up to 400 °C at 5° min− 1. The baseline was subtractedin all cases. Xerogel intermediates used for FT-IR and XRDanalyses were prepared in dynamic TG measurements up to thechosen temperature in the instrument itself. The thin films used forin situ spectroelectrochemical measurements were thermallytreated in a large muffle furnace since the substrate was tooheavy for heat-treatment in the TG instrument. It was not possibleto increase the temperature in themuffle furnace linearly with time.It was therefore set to a predetermined isothermal temperature atwhich the filmswere thermally treated for a certain time. IsothermalTG measurements of films deposited on platinum substrates wereperformed in advance to determine the degree of thermal de-composition of the films exposed to isothermal conditions. Theprocedure is described in detail in connection with Fig. 3. From theratio of the isothermal weight loss after a defined time and theweight loss associated with the decomposition of anhydrous nickelhydroxide, which for thin films was complete at 400 °C, the degreeof thermal decomposition could be estimated [2].

2.2.2. Evolved gas analysis (EGA)Thermogravimetry-mass spectroscopy (TG-MS) is a coupled

technique which enables determination of the nature and theamount of volatile product or products formed during thermaldecomposition [8]. TG-MS measurement of the xerogel wasperformed on a STA 409 Netzsch apparatus. Evolved gases weredetected using a Leybold Heraeus Quadrex 200 mass spectro-meter which was connected to the exhaust line of the TG instru-ment via a 75 cm long heated leak. During measurement thefurnace was purged with air (100 mL min− 1). The heating ratewas 5 K min− 1.

2.2.3. XRD analysisPowder diffraction data of xerogel sampleswere obtained using

a PANalytical X'Pert PRO diffractometer with Cu Kα radiationfrom 5 to 80 2θ in steps of 0.04 and a time per step of 1 s− 1.

2.2.4. FT-IR measurementsFourier transform infrared (FT-IR) spectroscopic measure-

ments were made using a Perkin Elmer System 2000 spectro-photometer with a resolution of 4 cm− 1. To obtain the transversaloptical modes, films were deposited on Si wafers. Thermal treat-ment of thin films heat-treated to different degrees was performedas a dynamic measurement using the TG instrument. For xerogelmeasurements KBr pellets embedding xerogels were prepared.Intermediates of xerogel and the thin film were heated to differenttemperatures which are marked in Fig. 1 with stars.

2.2.5. Spectroelectrochemical measurementsIn situ UV–VIS spectroelectrochemical measurements were

made in the range 360–1100 nm using a Perkin Elmer UV/VISLambda 2 Spectrometer, connected to an EG&G PARModel 273computer-controlled potentiostat–galvanostat. A potential scanrate of 10 mV s− 1 was used for CV measurements. A homemadethree-electrode spectroelectrochemical transmission cell, filledwith 40 mL of 0.1 M LiOH, was used. A Pt rod served as thecounter electrode, an Ag/AgCl/KClsat cell as the reference elec-trode (E=0.197 V) and a thin Ni-oxide film deposited on SnO2/F

Fig. 1. Dynamic TG curves of a thin film (left ordinate) and of the powder (rightordinate) in a dynamic air atmosphere. The initial weight of the thin film andsubstrate was 121.218 mg, while that of the powder was 9.880 mg. IR analysiswas performed for the intermediates marked in Fig. 1 with stars.

8265R. Cerc Korošec et al. / Thin Solid Films 516 (2008) 8264–8271


glass (square resistivity 13 Ω/ð) as the working electrode. Forbackground measurements, a cell filled only with electrolyte wasused.

2.2.6. Transmission electron microscopy (TEM) studyFor the TEM study a cross section of the sample on a bSiN

substrate was prepared using an adapted Gatan cross-sectionalTEM specimen preparation kit. After mechanical thinning anddimpling, ion milling using 3.8 keV argon ions at a 10° incidentangle was used. Samples were examined by a JEOL 2010 F trans-mission electron microscope, operated at 200 kV. The chemicalcomposition of the phases was determined using a Link ISIS-300energy dispersive X-ray spectroscopy (EDXS) system fromOxford Instruments with an ultra-thin window Si(Li) detector.Spectra were collected using a 200 keV electron beam with acollection time of 100 s. Spectra were collected in the energyrange up to 20 keV. During EDXS analysis special care was takento prevent beam-induced specimen damage, so we normally useda 50 um condenser aperture, spot size 3 and a beam diameterlarger than 20 nm.

3. Results and discussion

A comparison of dynamic TG measurements of a thin filmdeposited on a Pt foil, and the corresponding xerogel is presentedin Fig. 1. Comparison of the dynamic TG curve of the xerogel withthe published one [7] shows that there is some discrepancy inweight loss (Table 1), while the temperature range of the separatesteps is very similar. The reason for the discrepancy is probably thedifferent chemical compositions caused by the different precipita-tion solution (we used LiOH while the authors [7] used KOH) onthe one hand, and the lower moisture content in our sample on theother. However, there is a great difference in the course of thermaldecomposition of the thin film sample (dotted curve) with respectto that of the xerogel (solid curve). The first step in the weight losscurve is ascribed to dehydration of adsorbedwater and eliminationof intercalated water, and this occurred for both samples up to200 °C. Then the thermal decomposition of nickel hydroxideto nickel oxide began according to the reaction Ni(OH)2→NiO+H2O, which is evident from the second step of the TG curve.Nickel hydroxide exists in two polymorphic forms: crystallizedβ-Ni(OH)2 and α-Ni(OH)2, which is used to denominate a largeset of disordered Ni(II) hydroxides with low crystallinity andan excess of intersheet water and foreign ions [9]. During their

precipitation from aqueous salt solution, poorly crystallized α-type hydroxides or hydroxy-salts can be isolated. In our casean amorphous product was obtained, as evident from the XRDspectrumof the thermally untreated xerogel. Forβ-Ni(OH)2 it wasalready shown that from room temperature up to 250 °C dehy-dration takes place and above this temperature decompositionof Ni(OH)2 occurs [10,11].

Thermal decomposition of the thin film started at a tempera-ture 100 °C lower (220 °C) than that of the corresponding xerogel(325 °C). It is known that in dynamic measurements the decom-position temperature decreases with decrease in sample particlesize [12] since decomposition products first have to diffuse to thesurface of the solid particle before they evolve. This process takeslonger in the case of larger particles, while the temperature indynamic measurements is increasing. Therefore decompositiontemperatures for xerogel samples with an average size of someμm are shifted towards higher values in comparison with thinfilms with a thickness of some tens of a nm. On the basis of ourprevious examination this temperature difference was around30 °C [6], so we concluded that there should be another reason inthis case for such a large difference in the thermal stability of theprecipitated Ni(OH)x thin film with respect to the xerogel, mostprobably the different counter-ion. Lithium and sulfate ions arepresent in the structure due to the mode of preparation of thesample. Comparison of the IR spectra for the thin film and thexerogel showed that there is a large amount of sulfate ions in thexerogel sample, while in the thin film sample carbonate ions arepredominant. The origin of the carbonate ions is from atmosphericCO2 which chemisorbed on the basic surface after the thin filmhas been lifted from the LiOH solution. Reichle [13] reportedthat in layered hydroxides the order of affinity towards anions is:CO3

2−NSO42−, OH−NF−NCl−. During the ADD process the ex-

cess of negative sulfate ions (caused by carbonate insertion) pre-cipitated in the thin film must return to the solution due to theelectroneutrality condition. For the bulk material (xerogel) noevident insertion of carbonate ions took place. The reason is thatthe huge surface area of the amorphous Ni(OH)x precipitated onthe substrate acts as a nanostructured material. Many of the atomsare on the surface, allowing a nearly stoichiometric surface–gasreaction [14]. Carbonates are less thermally stable than sulfates.The mixture of NiCO3 and Ni(OH)2 thermally decomposed at

Table 1Comparison of weight loss between a xerogel prepared from KOH [7] and fromLiOH

Weight loss /%

1st step(25—approx.200 °C)

2nd step(200—approx.600 °C)

3rd step(600—approx.900 °C)

Xerogel [7],precip. agent KOH

15.8 17.9 8.9

Xerogel,precip. agent LiOH

10.7 11.7 12.7

Fig. 2. TG and EGA-MS curves of the xerogel.

8266 R. Cerc Korošec et al. / Thin Solid Films 516 (2008) 8264–8271


400 °C, while thermal decomposition of NiSO4 was completeat 900 °C [15]. Sulfate ions in the xerogel therefore stabilized theNi(OH)2 structure and decomposition began at higher tempera-ture. The third step in the TG curve of the xerogel above 600 °Cis a result of the decomposition of NiSO4 to NiO. On the TG–MScurve (Fig. 2) this is confirmed by the release of SO (mass peak48) and SO2 (mass peak 64). The mass peak of released water isshifted towards a higher temperature with regard to the derivativeTG (DTG) curve and its resolution is also poorer. The reason isthat the mass spectrometer is connected to the thermoanalyser viaa heated capillary 75 cm long and therefore a certain time shiftduring evolution of gas species and their detection in the massspectrometer is observed.

Fig. 3a and b presents the isothermal TG curve at 220 °C of athin film deposited on a Pt foil. In Fig. 3a the temperature profilein the furnace is presented on the right-hand axis and on the left-hand the weight of the film on the substrate in a percentage scale.The first step in the TG curve occurred during linear heating andis due to dehydration. Thermal decomposition of nickel hy-droxide to oxide occurs after the furnace reaches its isothermaltemperature and is fast in the beginning, but becomes slow withincreasing time. After 120 min the decomposition of anhydroushydroxide to oxide reaches 60%, as marked in Fig. 3b, whichrepresents the temperature dependence of the same data. Thetemperature in the furnace was increased again after 180 minat 220 °C to a final temperature of 400 °C, where decompositionto nickel oxide was complete (100%).

In Fig. 4a and b evolution of the XRD spectra of the xerogelwith increasing temperature is presented. The xerogel, heat-treated to 200 °C, showed the diffraction profile of an amorphousmaterial. The sharp peaks appearing at lower angles from approx-imately 15 to 35 2θ indicate the presence of LiSO4·H2O (PDF15-873) [16]. On heating to 400 °C, nickel oxide crystallites wereformed. The intermediate at 600 °C is a mixture of NiO, which isstill not well crystallised (marked in Fig. 4awith #, PDF 65-6920),two polymorphs of NiSO4

. 6H2O (PDF 47-1811 and 79-1654),anhydrous NiSO4 (PDF 72-1242) and LiSO4·H2O. At 900 °Ca mixture of crystallized nickel oxide and a small amount ofLiSO4·H2O is obtained (Fig. 4b). This means that sulfate ionscoordinated to nickel thermally decompose between 600 and800 °C, but those coordinated to lithium still remain in the struc-ture as lithium sulfate is thermally stable up to 860 °C [17]. Theidentified hydrated phases mean that between the TG and XRDmeasurements rehydration of the samples occurred.

The evolution of IR transmittance spectra during heat-treat-ment of the xerogel is shown in Fig. 5. The presence of watermolecules in the thermally untreated xerogel is revealed by abroad band positioned between 3800 and 2800 cm− 1, whichbelongs to the ν(O–H) stretching vibrations. The band is broad-ened due to hydrogen bonds. The band at 1630 cm− 1 is char-acteristic of the bending vibration of water. The intense peak

Fig. 3. Isothermal TG curves of a thin film deposited on a platinum foil at220 °C; time dependence (a) and temperature dependence (b).

Fig. 4. XRD spectra of xerogel intermediates: xerogel treated up to 200 °C—a;400 °C—b; 600 °C—c (a) and to 900 °C (b).



positioned at 1127 cm−1 corresponds to stretching vibrations offree sulfate ions with Td symmetry, meaning that a large amount ofsulfate ions are intercalated in the structure. The shoulder at1046 cm− 1 and splitting of the peak in the region of the bendingmodes of sulfate ions (686, 636 cm−1) are characteristic ofmonodentate coordination [18]. The ratio between the intensity ofthe peaks positioned at 1134 and 1051 cm−1 changed in thexerogel thermally treated up to 220 °C, so we assume that there aremore monodentately coordinated sulfate ions in this sample.On further thermal treatment, dehydroxylation of Ni(OH)2 occursleading to NiO formation. Up to 400 °C this process is notcomplete (Fig. 1) and the Ni–O stretching vibration (438 cm−1) isnot yet pronounced. Small NiO crystallites possess a large surfacearea. The band at 1727 cm− 1 is attributed to vibrations of thecarbonate ions that are bridge-bonded to the surface of NiO grains[19]. The band at 1278 cm−1 belongs to the combination ofstretching and bending vibrations of bidentately coordinatedcarbonate ions [20]. Most probably they originate from atmo-spheric CO2. The shoulder at 3580 cm

−1 indicates the presence ofnon-hydrogen bonded OH group. In the spectrum of the xerogelthermally treated to 600 °C, which is a mixture of several sulfates(see XRD spectrum, Fig. 4a), the symmetry of sulfate groups waslowered to C2ν, which corresponds to bidentate and bridge-bondedcoordination. The band for stretching Ni–O vibrations is thestrongest in the spectrum of the xerogel treated up to 900 °C.Sulfate ions are still present in the structure due to the presence ofLiSO4·H2O.

IR spectra of thin films deposited on silicon resin and thermallytreated to different extents are shown in Fig. 6. In the IR spectrumof the thermally untreated thin film the bands at 1488, 1419 and

861 cm− 1 represent stretching vibrations of the adsorbed car-bonate ion, which is planar and has D3h symmetry (ν2(CO3

2−)=879 cm− 1, ν3(CO3

2−)=1429 and 1492 cm− 1). The band at1588 cm− 1 arises from the stretching C–O vibrations of thebidentately coordinated carbonate ion [18]. The band at 667 cm− 1

arises from the bending vibration of δ(Ni–OH). After dehydrationthe shape of the spectrum resembles to the spectrum of a layeredturbostratic Ni(OH)2 [9], on which vibrations of carbonate ionsare superimposed. With a higher degree of thermal decomposi-tion, the band for stretching Ni–O vibrations becomes more andmore intense. The intensities of bands for carbonate ions dimin-ish during heat-treatment. Besides the intense Ni–O stretchingvibration, there are two peaks due to adsorbedmoisture and a peakof the silicon substrate (1088 cm− 1) in the spectrum of the filmthermally treated up to 400 °C.

In situ monochromatic optical transmittance changes(λ=480 nm) during chronocoulometric measurements of thinfilms thermally treated to different degrees are shown in Fig. 7.Fig. 7a represents the response at the beginning of the cyclingexperiment, while 7b shows the stabilised 101st cycle duringcharging at a potential of 0.6 V vs. Ag/AgCl (0–30 s) and dis-charging at 0.0 V vs. Ag/AgCl (30–60 s). At the beginning ofcycling, the highest degree of coloration was shown by the thinfilm where the degree of thermal decomposition from amor-phous anhydrous nickel hydroxide to oxide was 25% (120 minat 215 °C). After prolonged cycling the decrease of transmit-tance in its bleached state of 15% indicated that the colouring/bleaching process is not reversible. The change in transmittanceduring the 2nd coloration was only 15% for a thin film exposed

Fig. 5. IR transmittance spectra of xerogel intermediates: thermally untreatedxerogel—a; xerogel treated up to 220 °C—b; 400 °C—c; 600 °C—d and900 °C—e.

Fig. 6. IR transmittance spectra of thin films thermally treated to differentdegrees: thermally untreated film—a; film after dehydration (dynamically up to160 °C)—b; film with 45% decomposition (up to 255 °C)—c; 65%decomposition (up to 270 °C)—d and 100% decomposition (up to 400 °C)—e.



for 120 min at 220 °C (60% decomposition), but in the 101stcycle it reached a better value (43%) than the film with 25%decomposition (27%). After the 101st bleaching process, thedecrease in transmittance showed that reversibility was not yetachieved (Fig. 7b). An optimal optical response in the 101stcycle was exhibited by the film with 65% decomposition(120 min at 235 °C); the calculated coloration efficiency (CE)reached −37 cm2 C− 1, while at the beginning of cycling CEwas −32 cm2 C−1. For more thermally treated films the opticalmodulation was lower with regard to the optimized film. FromFig. 7b we can see that it took around 5 s to colour the film with25% decomposition and 90% decomposition, while films with60 and 65% decomposition needed more time (approx. 15 s).The reason probably lies in the different structures of these films.

The evolution of the cyclovoltammetric curves and the cor-responding in situ monochromatic transmittance changes ofthe optimized film (65% decomposition—120 min at 235 °C)are shown in Fig. 8a and b. This film exhibited excellent re-versibility in the 101st cycle. At the beginning of cycling the

basic form of the cyclovoltammetric curve was not yet developed(Fig. 8a). During the cycling process current densities becamelarger. The anodic and cathodic peaks became more and morepronounced, and the potential difference between them larger,indicating loosening of the structure. The changes in transmit-tance between the bleached and coloured state increased from13% in the 1st cycle to 45% in the100th cycle.

In Fig. 9 TEM micrographs of thin films thermally treatedfor 120 min at 235 °C (65% decomposition—a1, a2), for 24 hat 300 °C (b1, b2) and 24 h at 400 °C (c1, c2) are shown.We compared the microstructure of the optimized film with themicrostructure of a film which became electrochemically activeafter prolonged cycling in alkaline electrolyte and one whichstayed totally inactive [7]. The microstructure of the optimizedfilm is not aswell developed as in the case of the films prepared bythe sol–gel method [21]. It consists of particles (flakes) with a sizefrom 20 to 50 nm, and there is also some inner porosity. The openporosity of this sample enabled diffusion of LiOH during thecycling process through the whole thickness of the layer. Fromthe FT-IR spectrum it is evident that the film consists of two

Fig. 7. In situmonochromatic optical transmittance changes (λ=480 nm) duringchronocoulometric measurements of thin films thermally treated to differentdegrees (2nd cycle—(a), 101st cycle—(b)). The films were coloured at 0.6 V for30 s and bleached at 0.0 V for 30 s.

Fig. 8. Cyclovoltametric curves (a) and monochromatic transmittance changes(b) of the optimized film (65% decomposition—120 min at 235 °C).



structures: nickel hydroxide with inserted and bidentatelycoordinated carbonate ions and nickel oxide. The thickness ofthe film is around 300 nm (Fig. 9a1). The high-resolution TEM(HRTEM) image and the Selected Area Electron Diffraction(SAED) pattern (Fig. 9a2 with inset) of a cross section of the thinfilm show that the film is composed of an amorphous phase and afew nanometre (2–3 nm) sized crystalline particles. In the SAEDpattern,where thewell defined diffraction spots originate from thesilicon substrate, the faint and diffuse circle corresponds to the(002) plane of cubic NiO (marked with an arrow). Between the

silicon monocrystalline substrate and the NiO nanoparticles thereis a continuous amorphous layer around 2 nm thick, indicatingthat the NiO particles were formed during heat-treatment byhomogeneous nucleation. The estimated thickness of the film,annealed for 24 h at 300 °C (100% decomposition), is 130 nm(Fig. 9b1); that is more than twice as thin as the previouslydescribed film. Thatmeans that there is a considerable variation offilm thickness in the sample. The size of the nickel oxide grainsis 5 nm. This film was inactive at the beginning of cycling, butduring the cycling process became more and more active [7]. InFig. 9c1 (dark-field) and 9c2 (high-resolution) TEMmicrographsof the film which stayed inactive during the cycling experimentare shown. Fig. 9c1 (two different parts of the film sample areshown together) again confirms that there is a great variation offilm thickness, reaching 175 nm on one part and 350 nm on theother. Heat-treatment at 400 °C for 24 h resulted in the formationof larger grains of nickel oxide. The average diameter of the wellcrystallized (shaped) NiO grains is 10 nm, as can be seen from theimage on Fig. 9c2. The grains are embedded in an amorphousmatrix and overlap each other, producing characteristic Moirépatterns in the HRTEM image (bright and dark fringes inoverlapping areas).

4. Conclusion

Comparison of dynamic TG measurements made on ADDprepared thin films and the corresponding xerogels suggestedthat the two samples differ in more than sample size. The ob-served difference in onset decomposition temperature of nickelhydroxide was more than 100 °C. For sol–gel prepared samplesthe difference in onset decomposition temperature was around30 °C [6]. Besides, the xerogel sample exhibited a third step in thedynamic TG curve in the range from 600 to 800 °C, while the TGcurve of the thin film was flat in this range. With the help of FT-IRspectroscopy it was found that sulfate ions were replaced bycarbonate ions in thin film samples, most probably during theADD process when the sample was repeatedly pulled out of thesolution. At that time, atmospheric CO2 reacted with hydroxideions leading to carbonate formation. We conclude that all sulfateions were exchanged by carbonate ones since in the FT-IR spec-trum there is no evidence of sulfate vibrations. Sulfate ionsreturned to solution but we did not investigate the mechanism ofthis process. For the xerogel sample,whichwe prepared bymixingequal volumes of the two solutions, this process could not happen.

The film where decomposition of nickel hydroxide was 65%possessed an optimised electochromic response up to 101st cycle.From the FT-IR spectrum it is evident that it was composedof two structures: layered nickel hydroxide with inserted andbidentately coordinated carbonate ions and nickel oxide. TheTEM micrograph showed a thin film with a thickness of 300 nmwhich consisted of flakes 20–50 nm in size. Inside the flakes NiOnanoparticles (2–3 nm in diameter) were embedded in an amor-phous matrix (Fig. 9a1 and 9a2). Optical modulation of lessthermally treated films was better at the beginning of cycling, butpoorer in the 101st cycle with regard to the optimized film. Thebleaching process was not reversible, and the coloration lessintensive than for the optimized film. Optical modulation of more

Fig. 9. TEMmicrographs of the optimised film (65% decomposition: 120 min at235 °C—9a1 and 9a2), film exposed for 24 h to 300 °C—9b1 and 9b2 and for24 h to 400 °C—9c1 and 9c2.



thermally treated films (N65%) was lower with regard to theoptimized film (Fig. 7b). In our previous work we already noticedthat thermal decomposition of sol–gel prepared films whichcontain carbonate ions are always lower than in the case of sulfateions [6]. Carbonates probably hold the amorphous structuretightly together. Consequently they are hardly exchanged withOH− to obtain an electrochromically active Ni(OH)2 phase,leading to poorer response at the beginning of cycling. TEMshowed that the performance of films which consisted of onlyNiO phase (100% decomposition) depended on the grain size andcrystallinity. In the case when the size of the nanograins reached5 nm in diameter, the film became electrochemically/electro-chromically active during cycling. After 14 h the change intransmittance between the coloured and bleached state reached30% [7]. A film of 10 nm NiO grain size remained electro-chemically inert, probably due to its lower active surface size onone hand, and to the increase in crystallinity on the other.

References

[1] A. Azens, C.G. Granqvist, J. Solid State Electrochem. 7 (2003) 64.[2] R. Cerc Korošec, P. Bukovec, B. Pihlar, J. Padežnik Gomilšek, Thermochim.

Acta 402 (2003) 57.[3] A. Šurca, B. Orel, B. Pihlar, P. Bukovec, J. Electroanal. Chem. 408 (1996) 83.[4] X. Chen, X. Hu, J. Feng, Nanostruct. Mater. 6 (1994) 309.[5] R. Cerc Korošec, P. Bukovec, Thermochim. Acta 410 (2004) 65.[6] R. Cerc Korošec, P. Bukovec, Acta Chim. Slov. 53 (2006) 136.

[7] M.C.A. Fantini, G.H. Bezerra, in: C.M. Lampert, C.G. Granqvist (Eds.),Optical Materials Technology for Energy Efficiency and Solar EnergyConversion X, San Diego, U.S.A., July 25–26, 1991, SPIE Proceedings,vol. 536, 1991, p. 81.

[8] J. Mullens, in: M.E. Brown, P.K. Gallagher (Eds.), Handbook of ThermalAnalysis and Calorimetry, Principles and Practice, Vol. 1, Elsevier,Amsterdam, 1998, p. 509.

[9] P. Oliva, J. Leonardi, J.F. Laurent, C. Delmas, J.J. Braconnier, M. Figlarz,F. Filvet, J. Power Sources 8 (1982) 229.

[10] T.S. Horányi, Thermochim. Acta 137 (1989) 247.[11] Q. Song, Z. Tang, H. Guo, S.L.I. Chan, J. Power Sources 112 (2002) 428.[12] W.W. Wendlandt, Thermal Methods of Analysis, Interscience Publishers,

New York, 1964, p. 17.[13] W.T. Reichle, Solid State Ionics 22 (1986) 135.[14] K.J. Klabunde, R.S. Mulukutla, in: K.J. Klabunde (Ed.), Chemical and

Catalytic Aspects of Nanocrystals, Nanoscale Materials in Chemistry,Wiley, New York, 2001, p. 223.

[15] Atlas of Thermoanalytical Curves, in: G. Liptay (Ed.), Akadémiai KiadóBudapest, 1977, Vol. 5.

[16] Powder Diffraction File-4_Release, in: W.F. McClune (Ed.), InternationalCentre for Diffraction Data, Newtown Square, PA, 2006.

[17] P. Bukovec, N. Bukovec, in: Ž.D. Živkovič (Ed.), Thermal Analysis,Collection of Papers, Bor, Yugoslavia, 1984.

[18] K. Nakamoto, Infrared and Raman Spectra of Inorganic and CoordinationCompounds, 5th Ed.John Wiley & Sons, New York, 1997.

[19] G. Busca, V. Lorenzelli, Mater. Chem. 7 (1982) 89.[20] M. Figlarz, S. Le Bihan, C. R. Acad. Sci., 272 (1971) 580.[21] R. Cerc Korošec, P. Bukovec, B. Pihlar, A. Šurca Vuk, B. Orel, G. Dražič,

Solid State Ionics 165 (2003) 191.


Documents

Author's personal copy - Romana Cerc Korošec nickel oxide....pdf · Author's personal copy glass (square resistivity 13 /ð) as the working electrode. For background measurements,